Hyperglycemia Slows Embryonic Growth and SuppressesCell Cycle via Cyclin D1 and p21Devon E. Scott-Drechsel,

1Sandra Rugonyi,

1Daniel L. Marks,

2Kent L. Thornburg,

3

and Monica T. Hinds1

In pregnant women, the diabetic condition results in a three- tofivefold increased risk for fetal cardiac malformations as a resultof elevated glucose concentrations and the resultant osmoticstress in the developing embryo and fetus. Heart developmentbefore septation in the chick embryo was studied under twohyperglycemic conditions. Pulsed hyperglycemia induced bydaily administration of glucose during 3 days of developmentcaused daily spikes in plasma glucose concentration. In a secondmodel, sustained hyperglycemia was induced with a single in-jection of glucose into the yolk on day 0. The sustained modelraised the average plasma glucose concentration from 70 mg/dLto 180 mg/dL and led to decreased gene expression of glucosetransporter GLUT1. Both models of hyperglycemia reduced embryosize, increased mortality, and delayed development. Within the heartoutflow tract, reduced proliferation of myocardial and endocardialcells resulted from the sustained hyperglycemia and hyperosmolar-ity. The cell cycle inhibitor p21 was significantly increased, whereascyclin D1, a cell cycle promoter, decreased in sustained hypergly-cemia compared with controls. The evidence suggests that hyper-glycemia-induced developmental delays are associated with slowedcell cycle progression, leading to reduced cellular proliferation. Thesuppression of critical developmental steps may underlie thecardiac defects observed during late gestation under hyperglycemicconditions. Diabetes 62:234–242, 2013

Structural malformations of the heart affect 1% ofnewborn infants in the U.S. and are the leadingcause of noninfectious death among children.Maternal diabetes is associated with a three- to

five-fold increased risk for cardiac and valve malforma-tions, with hyperglycemia acting as a major teratogen. Thisis more alarming because of the rising incidence of di-abetes in the population. In 2010, an estimated 50 millionpeople worldwide had type 1 diabetes and 171 millionpeople had type 2 diabetes (1). An estimated 7% of U.S.pregnancies in 2010 had gestational diabetes, (2) resultingin 21% fetal mortality (3). Many pregnant women are un-aware of both their diabetes status during early pregnancyand the increased risk of morbidity and mortality that di-abetes has on their unborn child (4). Most pregnancies arenot recognized clinically until $2 weeks after conception,with the test for gestational diabetes typically performed at24–28 weeks. Frequently, diabetes is diagnosed and strict

glycemic control begins after the critical periods of em-bryogenesis and organogenesis. Thus, even blood glucosemanagement started in early pregnancy, but after organo-genesis, may be inadequate to prevent adverse pregnancyoutcomes (5). Although high maternal glucose concentra-tion and diabetes are associated with heart defects, thecellular and subcellular effects of maternal diabetes onembryonic heart development remain unknown.

Fetuses from diabetic mothers show a range of neuraltube (6) and cardiac malformations, including those of theoutflow tract (OFT) (7). To investigate the origins of OFTmalformations, we used a chick embryo model that iscommonly used to study heart development (8). In thismodel, hyperglycemia during organogenesis induced car-diac malformations that are apparent in late-stage fetuses(9). Glucose-induced malformations in animal models, in-cluding retarded growth and abnormal heart development(10), depend on the developmental stage of hyperglycemicexposure and glucose concentration (11). Previous studiesfound an increase in apoptosis and a decreased pro-liferation in the endothelial cells of 10-day-old chick em-bryos in which a 30% (wt/vol) glucose solution was injectedinto a chorioallantoic membrane vessel at day 8 (12). Em-bryonic stage delay as a result of maternal hyperglycemiahas been noted in mouse (11) and zebrafish (13). Previousstudies of pregnant mice with insulin-deficient streptozotocin-induced diabetes exhibited a developmental delay inpreimplantation embryos during their progression to ablastocyst stage both in vivo and in vitro (11). Additionally,in vitro cellular studies found that varying concentrationsof exogenous D-glucose (20 and 40 mmol/L) caused a de-crease in endothelial cell proliferation after 8 days of ex-posure (14). Embryonic cell viability has been shown todecrease when exposed to 33 mmol/L D-glucose for 96 h,whereas 44 mmol/L of D-glucose exposure for 24 h de-creased myoblast proliferation (15). A clinically relevantmodel of maternal hyperglycemia during cardiac organo-genesis, a critical stage of pregnancy, is needed. The cur-rent study examined developmental delay, stage delay, anddifferences between sustained and pulsatile glycemicstress on cardiac development.

Using chick embryos, we developed and tested two in vivomodels of hyperglycemia: 1) sustained hyperglycemia, mim-icking the exposure of embryos of pregnant women withpoor glycemic control, and 2) pulsatile hyperglycemia, rep-resenting embryonic conditions of diabetic mothers aftermeal-associated increases in glucose. The models tested thehypothesis that hyperglycemia slows embryonic developmentthrough an increase in osmotic stress, a decrease of glucosetransport, and cell cycle suppression. We investigated theunderlying causes of developmental delay and growth sup-pression in the chick embryo by examining cell cycle regu-lators p21 and cyclin D1, the glucose transport gene GLUT1,and OFT myocardial and endocardial cell proliferation.

From the 1Biomedical Engineering Department, Oregon Health & Science Uni-versity, Portland, Oregon; the 2Papé Family Pediatric Research Institute, Or-egon Health & Science University, Portland, Oregon; and the 3Heart ResearchCenter, Oregon Health & Science University, Portland, Oregon.

Corresponding author: Monica T. Hinds, hindsm@ohsu.edu.Received 10 February 2012 and accepted 14 July 2012.DOI: 10.2337/db12-0161� 2013 by the American Diabetes Association. Readers may use this article as

long as the work is properly cited, the use is educational and not for profit,and the work is not altered. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

See accompanying commentary, p. 27.

234 DIABETES, VOL. 62, JANUARY 2013 diabetes.diabetesjournals.org

ORIGINAL ARTICLE

RESEARCH DESIGN AND METHODS

Preparation of chick embryos. The chick embryo was used as a modelsystem of cardiac development for several reasons: 1) Early development ofhuman hearts is similar to that of chick embryos but with a faster time course(21 days in chicks compared with 9 months in humans) (16); 2) congenitalheart defects have been well documented in the chick embryo (17,18); 3) thechick embryo is transparent, easy to access, and easily observed without af-fecting cardiac function (19); and 4) the chick OFT is a key structure thatconnects the ventricle to the arterial system and is the site where a number ofserious septal and valve anomalies originate in human hearts (20).

A randommix of nonincubated, fertilized, White Leghorn chicken eggs wereobtained from a hatchery (Featherland Farms Inc., Coburg, Oregon). Twomodels of hyperglycemia, pulsed and sustained, were developed. To inducea pulsed elevation in plasma glucose concentration, a 2-cm window was madein the blunt end of fertilized chicken eggs before incubation and 50 mL of 30 or50 mmol/L D-glucose, L-glucose (osmotic control), or chick Ringer saline so-lution (vehicle control) was dropped in the egg air sac directly on top of theembryo. Drops were added on days 0, 1, 2, and 3 of incubation. For a sustainedelevation in plasma glucose concentration, a 2-cm window was made abovethe air sac of the egg and using a 30-gauge needle, 600 mL of D-glucose in 0.7%saline was slowly injected directly into the egg yolk before incubation; carewas taken to avoid injecting bubbles and damaging the blastocyst. To establisha clinically relevant elevation of plasma glucose, D-glucose was added atconcentrations of 30, 50, 100, 250, 500, and 750 mmol/L. Control embryosreceived an injection of L-glucose or vehicle. After yolk or air sac injections, allchicken eggs were sealed with plastic wrap and incubated at 38°C with theblunt end up in a humid atmosphere.Plasma glucose measurements. Plasma glucose measurements were takenwhen embryos reached Hamburger-Hamilton developmental stage 24 (HH24)(21). Using a dissecting microscope, the ventricle was punctured with a 30-gauge needle; blood was withdrawn into 3.8% sodium citrate and stored on ice.At this early embryonic age, 30–40 mL of blood was obtained from each HH24embryo. The blood was centrifuged (2,500 rpm for 15 min) to isolate plasma.The plasma was frozen at 280°C until biochemical analyses were performed.Plasma glucose concentrations were measured using a glucose assay kit(Sigma) according to the manufacturer’s instructions. Briefly, glucose oxidase/peroxidase reagent was added to the isolated plasma and incubated for 30 minat 37°C. The addition of N2SO4 stopped the reaction. Using a spectrophotom-eter set at 540 nm, the absorbance of each solution was measured againsta standard curve to measure glucose concentration.Viability and growth measurements. To assess the physiological effects ofsustained and pulsed hyperglycemia on the chick embryos, viability and growthmeasurements were made. After 96 h of incubation, the embryonic de-velopmental stage was determined based on developmental markers and theHamburger-Hamilton chick embryonic developmental scale (21). Embryoswere removed from their eggs and weighed to quantify differences in em-bryonic size. Percent mortality was measured by counting the number of livingversus dead embryos at 96 h.Proliferation. Proliferating cells in the OFT were DNA labeled with a Click-iTEdU kit (Invitrogen) by detection of a thymidine analog in the intact double-stranded DNA. Four hundred microliters of 10 mmol/L EdU antibody diluted inchick saline (0.75 g/L NaCl), was dropped on each living HH24 embryo throughthe egg window and incubated for 3 h. The embryos were removed and fixedin 4% paraformaldehyde overnight. After fixation, the OFTs were dissectedfrom the embryonic hearts, cut longitudinally, and stained according to themanufacturer’s instructions. Briefly, OFTs were permeabilized in Triton X-100and incubated in the reaction cocktail (reaction buffer, CuSO4, Alexa Fluorazide, and buffer additive) for 30 min. The OFTs were blocked with 1% BSAand incubated in anti-myosin heavy chain antibody (1:10) (MF20; De-velopmental Studies Hybridoma Bank) for 1 h. The OFTs were incubated insecondary antibody (1:100 goat anti-mouse IgG Alexa Fluor 564; Invitrogen)for 1 h. The DNA nuclear counterstain Hoechst (1:1000) was applied followedby mounting in SlowFade (Invitrogen). The stained OFTs were imaged usingconfocal laser scanning microscopy (Bio-Rad MRC500 on a Zeiss Axiovertinverted microscope). To analyze proliferation in all layers of the OFT wall,images were collected in three dimensions as a z-series of ;3 mm per step.The z-series was analyzed using ImageJ (National Institutes of Health) soft-ware and reconstructed as a three-dimensional map of the OFT (22–24);images were taken at five sections along the length of the OFT from theproximal to the distal end. Proliferating cells were counted in each section ofthe z-series to get a ratio of proliferating cells versus nonproliferating cells.Quantitative PCR. For quantitative PCR analysis, the dissected OFTs fromHH24 embryos were placed in a lysate solution and homogenized. Total RNAwas isolated using an RNeasy kit (Qiagen) and reverse-transcribed usingSuperScript III Reverse Transcriptase (Invitrogen). Quantitative PCR wasperformed with a 96-well Agilent Mx3005P QPCR System and Power SYBR

Green. Forward and reverse primers (spanning at least 2 exons) were madeusing Ensembl genome browser and Primer3 software. Gene expression wasdetermined relative to a calibrator, which comprised pooled OFTs from un-opened and untreated eggs at HH24, and normalized to the housekeeping genePPIA using the standard curve method and taking primer efficiencies intoaccount. Forward and reverse primers for p21, cyclin D1, GLUT1, and PPIAwere as follows:

1. p21: TCCTCCTCCTACCAGAGATG, TGTACCTGAGGCTCCTTGTC;2. cyclin D1: TCGGTGTCCTACTTCAAGTG, GGAGTTGTCGGTGTAAATGC;3. GLUT1: TCTCTGTCGCCATCTTCTCG, TGGTGAGGCCAGAATACAGG; and4. PPIA: TGACAAGGTGCCCATAACAG, TTCTCGTCGGCAAACTTCTC.

Protein analysis. Western blot protein quantification was performed to de-termine whether the transcription and protein levels correlated for p21 andcyclin D1. Protein was extracted from three pooled OFTs using a lysis solutionof 5 mmol/L Tris-HCl, 5 mmol/L EGTA, 5 mmol/L EDTA, and 0.06% SDS (RIPAbuffer; Millipore, Temecula, CA) and protease inhibitor (Roche, Indianapolis,IN) and was quantified using a bicinchoninic acid protein assay (Pierce).Protein (15 mg) was separated by SDS-PAGE and transferred onto reinforcednitrocellulose membranes before incubation with primary antibodies, mousemonoclonal p21, mouse monoclonal cyclin D1, and mouse monoclonala-tubulin (Santa Cruz Biotechnology) overnight followed by incubation withhorseradish peroxidase goat anti-mouse IgG (Cell Signaling). Protein was vi-sualized by chemiluminescence (SuperSignal; Pierce), and digitally quantified(ImageJ) to detect endogenous levels of total protein with respect to a-tubulin.Statistical analysis. Data are presented as mean6 SD. The effects of glucosetreatment were analyzed by ANOVA with Fisher least significant difference(LSD) post hoc. Results were considered significant at P , 0.05.

RESULTS

Plasma glucose. To determine whether the addition ofexogenous glucose increased the blood glucose concen-tration in embryonic chicks, plasma glucose concen-trations were measured. Normal untreated embryos hada plasma glucose concentration of 70 6 15 mg/dL. Thesingle addition of 50 mmol/L D-glucose into the air sac onday 4 (HH24) into a previously unopened egg led to a steeptransient increase in circulating plasma glucose concen-tration, which peaked at 176 6 20 mg/dL at 60 min andreturned to a normal concentration of 76 6 15 mg/dL by120 min (Fig. 1A). The 30 mmol/L L- and D-glucose–treatedembryos followed the same pattern. Repeating the glucoseaddition into the air sac on days 0, 1, 2, and 3 of incubationdid not alter the blood glucose level on day 4 (data notshown). To induce sustained hyperglycemia, a higherconcentration of glucose with continuous delivery wasneeded. Glucose additions of 30–500 mmol/L to the yolkdid not raise plasma glucose concentrations to a clinicallysignificant level (Fig. 1B). The injection of 750 mmol/LD-glucose raised plasma glucose concentration in HH24embryos to 177 6 20 mg/dL, a clinically significant levelcompared with 72 6 9 mg/dL in the stage-matched un-opened chick embryos (Fig. 1B). To determine whetherthe sustained hyperglycemic state was maintained later indevelopment, plasma glucose concentrations of D-glucose-injected embryos on day 6 of incubation were measuredand found to be 165 6 30 mg/dL compared with vehicle-treated embryos with a plasma glucose concentration of80 6 8 mg/dL. Determination of the expression of GLUT1was used to evaluate the role of glucose transportersduring hyperglycemia. The gene expression level ofGLUT1 (Fig. 1C) was decreased by twofold for D-glucose-treated embryos compared with vehicle.Viability and growth. By measuring the viability and em-bryonic growth of treated and untreated embryos at 96 h, theeffects of exogenous glucose on embryonic developmentwere determined. Pulsed hyperglycemia caused a small de-velopmental stage delay, with an average embryonic age of

D.E. SCOTT-DRECHSEL AND ASSOCIATES

diabetes.diabetesjournals.org DIABETES, VOL. 62, JANUARY 2013 235

HH23 or a 4- to 6-h delay in their development comparedwith vehicle controls at HH24 (Table 1). A decreasing trendin embryonic weight was found in these pulsed hyperglyce-mic embryos because the average weights of D- and L-glucose-treated embryos decreased 20 and 10%, respectively,compared with vehicle controls (Table 1). Pulsed D-glucoseembryos had 10% mortality, whereas L-glucose– and vehicle-treated embryos had 7 and 1% mortality, respectively.

Although the pulsed hyperglycemic condition resulted ina slight reduction in growth and embryo viability, thesustained hyperglycemic model induced larger changes inthe embryos. Sustained hyperglycemia produced a stagedelay, equivalent to a 12-h delay in development, in theembryos, with an average stage of HH22 compared withthat for the L-glucose and vehicle control embryos of HH23and HH24, respectively (Table 1). The average weight of

FIG. 1. The addition of exogenous glucose to chick embryos increased plasma glucose concentration. A: Transient plasma glucose concentration inchick embryos at day 4 (HH24) for embryos treated with a single pulse of 30 mmol/L D-glucose, 50 mmol/L D-glucose, or 30 mmol/L L-glucose with anenlargement of the critical area of glucose spike for 0–90 min. B: Plasma glucose concentration for HH24 embryos after yolk injection of varyingconcentrations of D-glucose at day 0. A clinically significant concentration of plasma glucose resulted from the 750 mmol/L D-glucose injection. C:Under sustained hyperglycemia, expression of GLUT1 levels in the OFT for vehicle-, L-glucose–, and D-glucose–treated embryos. Data are mean 6SD with trend lines for n = 10. *Difference from vehicle (P < 0.05) using ANOVA with Fisher LSD post hoc.

HYPERGLYCEMIA SUPPRESSES EMBRYONIC DEVELOPMENT

236 DIABETES, VOL. 62, JANUARY 2013 diabetes.diabetesjournals.org

D- and L-glucose–treated embryos decreased by 50 and10%, respectively, compared with vehicle controls. Percentmortality was high for all the yolk-injected embryos be-cause the insertion of a syringe alone into the yolk of theeggs causes a 50% mortality rate for embryos (25). Usingvehicle-injected embryos as a baseline, the sustained hy-perglycemic condition significantly increased mortality by30% compared with vehicle control and 11% comparedwith L-glucose.OFT cellular proliferation. A decrease in size and stagedelay was found in both pulsed and sustained hypergly-cemic embryos. To examine a potential cause of this delay,the degree of cell proliferation in stage-matched HH24OFTs was determined. There were no significant differ-ences in the number of proliferating endocardial or myo-cardial cells in the OFT for the pulsed model (Fig. 2).Sustained hyperglycemia resulted in a 10% average de-crease in OFT cell proliferation for endocardial and myo-cardial cells for both the D- and L-glucose–treated embryoscompared with the vehicle controls (Fig. 3). Proliferationdid not vary with location in the OFT.Cell cycle–regulated gene and protein expression. Theexpression levels of p21 and cyclin D1 in the OFT wereevaluated in sustained hyperglycemic embryos to de-termine whether cell cycle dysregulation contributed tothe reduction of cell proliferation and embryonic de-velopmental delay. Gene expression for p21 significantlyincreased by twofold in D-glucose–treated embryos,whereas no significant difference was found between thevehicle and L-glucose–treated embryos (Fig. 4A). The gene

expression level of cyclin D1 (Fig. 4B) was decreasedby threefold for D-glucose–treated embryos comparedwith vehicle and L-glucose control embryos. Similarly, thep21 protein expression was significantly increased forD-glucose–treated embryos, and there was a trend towarddecreased protein expression for cyclin D1 in D-glucose-treated embryos compared with vehicle and L-glucosecontrols (Fig. 5).

DISCUSSION

The offspring of mothers with unrecognized type 1, type 2,or gestational diabetes have a high risk for being born withcongenital anomalies, macrosomia, and neonatal, child-hood, and adult complications (5). Congenital malforma-tions found in the children of diabetic mothers includeneural tube defects, which can lead to anencephaly, spinabifida aperta, meningocele, and encephalocele, and car-diovascular defects, such as septal defects and malfor-mations of the heart and great vessels (26). Although thesecomplications are known, few models exist to determinethe effects and underlying causes of varying concen-trations of glycemic stress on the development of con-genital malformations during organogenesis. Thus, wehypothesized that hyperglycemia slows embryonic de-velopment through an increase in osmotic stress and thesuppression of the glucose transporters and the cell cycle.To test this hypothesis, we developed and characterizedtwo in vivo embryonic chick models to mimic embryonicconditions during maternal hyperglycemia: 1) pulsed hy-perglycemia (associated with meal-induced hyperglyce-mia) and 2) sustained hyperglycemia (associated withpoorly controlled maternal hyperglycemia). The embryosin this study were exposed to either a pulse or a sustainedincrease in blood glucose concentration. The plasma glu-cose concentrations in the pulsed hyperglycemic modelwere characterized by a peak at 60 min with a return tonormal by 120 min. Although measurements at earlier de-velopmental stages were not possible, we speculate thateach daily glucose treatment caused a temporary spike inembryonic plasma glucose concentration followed bya return to normal. A longer-term exposure to glucosepulses (injections on days 0, 1, 2, and 3 of incubation),which has previously been shown at day 18 to both resultin a significant increase in plasma glucose and lead toembryonic malformation (28), did not result in increasedcirculating glucose concentrations at day 4. An increase inthe plasma glucose concentration of HH24 chicks at day 4was accomplished in the sustained hyperglycemic model,which maintained an elevated blood glucose concentrationuntil at least day 6 of development. In both models, theaddition of D- and L-glucose resulted in clinically relevantplasma glucose concentrations in the chick embryos, ele-vating the average circulating glucose concentration to 180mg/dL compared with the normal plasma glucose level of70 mg/dL. To create these high concentrations, largequantities of exogenous glucose were added as follows: 50and 750 mmol/L D-glucose in the pulse and sustainedmodels, respectively, which affected the osmolarity of theblood. Hyperglycemic conditions in humans and in thepresent model induce osmotic stress through increasedactivity in the polyol pathway and excess generation ofsorbitol (27). Thus, L-glucose was used as an osmoticcontrol because cells do not readily uptake L-glucose,whereas glucose transporters such as GLUT1 facilitatethe transport of D-glucose across the cell membrane. The

TABLE 1The effects of pulsed and sustained hyperglycemia on 96-h chickembryos

Treatment

Bloodglucose(mg/dL)(n = 10)

Stage(n = 10)

Embryoweight(mg)

(n = 10)

Mortality(%)

(n = 10)

Unopened 75 6 12 HH24 110 6 13 —

Pulsedhyperglycemia

Vehicle 76 6 15 82% HH2418% HH23 103 6 20 1

L-glucose 74 6 12 46% HH2454% HH23 95 6 23 7*

D-glucose 80 6 949% HH2424% HH2327% HH22

80 6 20 10*

Sustainedhyperglycemia

Vehicle 72 6 976% HH2439% HH238% HH22

126 6 17 50

L-glucose 120 6 15*44% HH2438% HH2318% HH22

85 6 22 60*

D-glucose 177 6 20*

20% HH2430% HH2343% HH227% , HH22

65 6 15* 71*

Data are mean 6 SD unless otherwise indicated. Effects are bloodglucose concentration, developmental stage with percentage of em-bryos per stage, embryo weight, and percent embryonic mortality forvehicle (saline), L-glucose, and D-glucose treatments. *Significant dif-ference from vehicle (P , 0.05).

D.E. SCOTT-DRECHSEL AND ASSOCIATES

diabetes.diabetesjournals.org DIABETES, VOL. 62, JANUARY 2013 237

L-glucose controls enabled us to distinguish between theeffects of hyperglycemia and hyperosmolarity.

Glucose is essential for normal fetal growth and me-tabolism, but excess amounts are detrimental to fetal de-velopment. Facilitated glucose transport across the plasmamembrane of a cell is mediated by glucose transporters.The distribution of glucose transporters varies in organ

systems based on differing tissue requirements for glu-cose. GLUT1 is the most prominent glucose transporter inembryonic development, particularly in the embryonicchick heart, yet it is insulin independent (28). GLUT4 is aninsulin-dependent glucose transporter in human hearts(28), but it is not present in the avian population, and theexistence of an insulin-dependent glucose transporter in

FIG. 2. Confocal images indicating the proliferating cells for the pulsed hyperglycemic condition. The nuclei of the proliferating endocardial cellsfor the vehicle-treated (A), L-glucose–treated (B), and D-glucose–treated (C) OFTs are indicated by the green stain. Myocardial cell proliferationfor vehicle (D), L-glucose (E), and D-glucose (F) is designated by the green nuclear stain (Alexa Fluor 488) with the red myosin stain (Alexa Fluor564) to differentiate the cell type. Original magnification 4003. Scale bar represents 20 mm. The percentage of volumetric (G), endocardial cell(H), and myocardial cell (I) proliferation in the OFTs of HH24 embryos for pulsed additions of vehicle, L-glucose, and D-glucose are shown. Data aremean 6 SD for n = 10.

HYPERGLYCEMIA SUPPRESSES EMBRYONIC DEVELOPMENT

238 DIABETES, VOL. 62, JANUARY 2013 diabetes.diabetesjournals.org

birds remains unknown. In the current study, sustainedhyperglycemic conditions resulted in a decrease in geneexpression of GLUT1 in the OFT. Similarly, in vivo pre-implantation embryo studies with mouse models haveshown a decrease in GLUT1, GLUT2, and GLUT3 as a re-sult of maternal hyperglycemia (29). The downregulationof GLUT1 as a result of maternal hyperglycemia could leadto low levels of intracellular glucose, depriving cells of

nutrition and predisposing embryos to developmentaldelays and other malformations (30). GLUT1-deficent em-bryonic mice have been shown to slow growth, reducebody weight, and lead to other malformations (31) similarto that of the hyperglycemic chick embryos. Therefore, thedownregulation of GLUT1 in the current model facilitatesexcess amounts of circulating glucose while reducing in-tracellular glucose concentrations, which may play a role

FIG. 3. Confocal images indicated the proliferating cells for the sustained hyperglycemic condition. The nuclei of the proliferating endocardial cellsfor the vehicle-treated (A), L-glucose–treated (B), and D-glucose–treated (C) OFTs are indicated by the green stain. Myocardial cell proliferationfor vehicle (D), L-glucose (E), and D-glucose (F) is designated by the green nuclear stain (Alexa Fluor 488) with the red myosin stain (Alexa Fluor564) to differentiate the cell type. Original magnification 4003. Scale bar represents 20 mm. The percentage of volumetric (G), endocardial cell(H), and myocardial cell (I) proliferation in the OFTs of HH24 embryos for sustained additions of vehicle, L-glucose, and D-glucose are shown. Dataare mean 6 SD for n = 10. *Difference from vehicle (P < 0.05) using ANOVA with Fisher LSD post hoc.

D.E. SCOTT-DRECHSEL AND ASSOCIATES

diabetes.diabetesjournals.org DIABETES, VOL. 62, JANUARY 2013 239

in embryonic malformations from maternal diabetes-inducedhyperglycemia.

Embryonic stage delay as a result of hyperglycemia hasbeen demonstrated in zebrafish (13), mice (11), and rats(32). Both models in the current study showed a de-velopmental delay, increased embryonic mortality, anda smaller embryo size, similar to fetuses of diabeticmothers during early pregnancy (33). However, sustainedhyperglycemia had a more pronounced effect on embryonicdevelopment. These outcomes are in agreement with othersthat have shown an increase in embryo mortality and ab-normal growth as a result of a concentration-dependenteffect of D-glucose (34). Osmotic stresses or a pre-dominance of female embryos (male chicks hatch earlierthan female chicks [35]) are possible explanations for thestage delay and decrease in weight of L-glucose embryos.Thus, a random sample of both male and female embryoswas used. The high concentration of circulating glucoseand osmotic stress in the embryo that accompany mater-nal diabetes results in embryonic delay and impairs allorgan systems (36,37), particularly the cardiovascularsystem. Therefore, we examined the mechanisms of thisdelay in the heart OFT to determine whether cellulardysfunction in the OFT contributes to the development ofcardiac malformations.

In the OFT, a decrease in proliferation of both endo-cardial and myocardial cells was demonstrated in thesustained hyperglycemic model. The development ofthe embryonic heart and OFT depends on the functions ofthe endocardial, myocardial, and neural crest cells. Re-duced proliferation in neural crest cells has been shown

under conditions of high glucose concentrations in bothcells cultured in 50 mmol/L D-glucose and mouse embryoswith streptozotocin-induced diabetes (38,39). In the cur-rent model, neural crest cells are likely a factor in thehyperglycemia-induced OFT changes from the onset ofincubation; however, the effect of hyperglycemia on neuralcrest cell proliferation remains to be determined in thismodel. The decrease in cellular proliferation in the OFTmay contribute to OFT underdevelopment and producea malformed structure. Zhao (40) found a decrease in mi-tosis in the endocardial cushions of the OFT in embryos ofdiabetic mice, changing the structure of the cushions.Thus, the decrease in proliferation may be instrumental inthe formation of similar malformations seen at late stagesof human development, such as semilunar valve malfor-mations. The proliferation for both endocardial and myo-cardial cells significantly decreased for both D- andL-glucose–treated embryos, suggesting a role for osmoticstress in the reduction of cell proliferation. This finding isin agreement with others who have shown decreased en-dothelial cell proliferation because of high glucose con-centration and hyperosmolarity (41). Because a significantdecrease in proliferation was found in the sustained hy-perglycemic case, cell cycle gene signaling was in-vestigated.

In the current study, we found that regulators of the cellcycle were altered in cardiac cells located in the OFT wallof hyperglycemic embryonic chick hearts. Cell proliferationis the end point of an ordered set of phases in the cell cycle,which are regulated by cyclins and cyclin-dependent ki-nases. Cyclins drive cells from the G1 phase to the S phase

FIG. 4. Sustained hyperglycemia caused changes in expression of genes regulating the cell cycle. The fold changes of gene expression werecompared with an OFT from an unopened calibrator for p21 (A) and cyclin D1 (B) for the embryo yolk injected with vehicle, L-glucose, or D-glucose.Data are mean 6 SD for n = 10. *Difference from vehicle (P < 0.05) and †difference from L-glucose (P < 0.05) using ANOVA with Fisher LSD posthoc.

HYPERGLYCEMIA SUPPRESSES EMBRYONIC DEVELOPMENT

240 DIABETES, VOL. 62, JANUARY 2013 diabetes.diabetesjournals.org

and from the G2 phase to the mitosis phase by binding tocyclin-dependent kinases at critical points during the cellcycle. This binding leads to the phosphorylation of targetproteins, which enables progression to the next phase. Thecyclin-dependent kinase inhibitor p21 is associated withcell cycle arrest in the G1 phase and, thus, is a criticalmediator of growth arrest (42) by inhibiting mitosis, DNAreplication, and repair (43). During the cell cycle, the G1/Sphase depends on cyclin D1 and is negatively regulated byp21. Previous studies in an ovine model of cardiomyocytesdemonstrated that p21 and cyclin D1 are reliable indica-tors of cell proliferation (44,45), with an increase in p21and a decrease in cyclin D1 contributing to reduced cellproliferation. Under sustained hyperglycemic conditions,embryonic chick cardiac cells may have arrested in the G1

phase because of the suppression of cyclin D1 and theupregulation of p21, which would lead to a reduction ofendocardial and myocardial cell proliferation. Althoughthe decrease in cardiac cell proliferation was similar in theOFTs of embryos treated with D- and L-glucose, there weresignificant differences in the expression of p21 and cyclinD1 between the D-glucose–treated embryos and theL-glucose controls. High osmolarity can also contribute todecreased cell proliferation by arresting cells in their G1/Sand G2 phases. For example, hyperosmolarity can activatethe G2/M checkpoint through p38 activation, which causesa drop in the cyclin-dependent kinase 2 activity (46). Thus,the inhibition of cell proliferation that we found in theD- and L-glucose–treated embryos may be through differentmechanisms of cell cycle arrest or another mechanismassociated with proliferation. Although not examined here,apoptosis is another potential mechanism leading to ma-ternal diabetes-induced stage and developmental delay.Hyperglycemia can lead to apoptosis during embryogene-sis (47). The apoptotic pathway may be triggered by re-active oxygen species generation, and tissue damage fromthis apoptosis could lead to abnormalities (48). However,more research needs to be conducted on this pathway.

Maternal diabetes has detrimental effects on embryos,increasing the rates of embryonic malformations. In thisavian model, elevated glucose decreased the glucosetransporter GLUT1 and resulted in embryonic stage delay,development retardation, and increased embryonic mortal-ity. Additionally, sustained hyperglycemia reduced endo-cardial and myocardial cell proliferation if the heart OFT,which may be the associated with the downregulation of thecell cycle by p21 and cyclin D1. This change in cellularproliferation is likely to affect the development of the OFTand contribute to malformations later in development. Im-portantly, this model enables the study of concentration-dependent effects of hyperglycemia on cardiac developmentduring early organogenesis and determination of the un-derlying mechanisms of hyperglycemia-induced cardiacmalformations.

ACKNOWLEDGMENTS

A portion of this work was funded by National Institutes ofHealth Grant R01-HL-094570.

No potential conflicts of interest relevant to this articlewere reported.

The funder played no role in the conduct of the study,collection of data, management of the study, analysis of data,interpretation of data, or preparation of the manuscript.

D.E.S.-D. conducted experiments, researched data, andwrote the manuscript. S.R. and K.L.T. contributed to discus-sion and reviewed and edited the manuscript. D.L.M. con-tributed to discussion and reviewed the manuscript. M.T.H.researched data, contributed to discussion, and reviewed andedited the manuscript. D.E.S.-D. is the guarantor of this workand, as such, had full access to all the data in the study andtakes responsibility for the integrity of the data and theaccuracy of the data analysis.

The authors thank Natasha Chattergoon (Oregon Health& Science University, Heart Research Center) for provid-ing assistance with the Western blot protein analysis.

REFERENCES

1. Roman MA. Preconception care for women with preexisting type 2 di-abetes. Clin Diabetes 2011;29:10–16

FIG. 5. Protein levels of p21 and cyclin D1 in sustained hyperglycemicOFTs. p21 significantly increased in D-glucose–treated OFTs, whereascyclin D1 trended toward decreased protein expression compared withL-glucose and vehicle. Data are mean 6 SD for n = 5. *Difference fromvehicle (P < 0.05) and †difference from vehicle (P < 0.05) usingANOVA with Fisher LSD post hoc.

D.E. SCOTT-DRECHSEL AND ASSOCIATES

diabetes.diabetesjournals.org DIABETES, VOL. 62, JANUARY 2013 241

2. Chu SY, Callaghan WM, Kim SY, et al. Maternal obesity and risk of ges-tational diabetes mellitus. Diabetes Care 2007;30:2070–2076

3. Madri JA, Enciso J, Pinter E. Maternal diabetes: effects on embryonicvascular development—a vascular endothelial growth factor-A-mediatedprocess. Pediatr Dev Pathol 2003;6:334–341

4. Evers IM, de Valk HW, Visser GH. Risk of complications of pregnancy inwomen with type 1 diabetes: nationwide prospective study in The Neth-erlands. BMJ 2004;328:915

5. Eriksson UJ. Congenital anomalies in diabetic pregnancy. Semin FetalNeonatal Med 2009;14:85–93

6. Eriksson UJ, Borg LA, Cederberg J, et al. Pathogenesis of diabetes-inducedcongenital malformations. Ups J Med Sci 2000;105:53–84

7. Loffredo CA, Wilson PD, Ferencz C. Maternal diabetes: an independentrisk factor for major cardiovascular malformations with increased mor-tality of affected infants. Teratology 2001;64:98–106

8. Li P, Yin X, Shi L, Liu A, Rugonyi S, Wang RK. Measurement of strain andstrain rate in embryonic chick heart in vivo using spectral domain opticalcoherence tomography. IEEE Trans Biomed Eng 2011;58:2333–2338

9. Hughes AF, Freeman RB, Fadem T. The teratogenic effects of sugars onthe chick embryo. J Embryol Exp Morphol 1974;32:661–674

10. Roest PA, van Iperen L, Vis S, et al. Exposure of neural crest cells to el-evated glucose leads to congenital heart defects, an effect that can beprevented by N-acetylcysteine. Birth Defects Res A Clin Mol Teratol 2007;79:231–235

11. Moley KH, Chi MM, Manchester JK, McDougal DB, Lowry OH. Alterationsof intraembryonic metabolites in preimplantation mouse embryos exposedto elevated concentrations of glucose: a metabolic explanation for thedevelopmental retardation seen in preimplantation embryos from diabeticanimals. Biol Reprod 1996;54:1209–1216

12. Larger E, Marre M, Corvol P, Gasc JM. Hyperglycemia-induced defects inangiogenesis in the chicken chorioallantoic membrane model. Diabetes2004;53:752–761

13. Liang J, Gui Y, Wang W, Gao S, Li J, Song H. Elevated glucose inducescongenital heart defects by altering the expression of tbx5, tbx20, and has2in developing zebrafish embryos. Birth Defects Res A Clin Mol Teratol2010;88:480–486

14. Varma S, Lal BK, Zheng R, et al. Hyperglycemia alters PI3k and Akt sig-naling and leads to endothelial cell proliferative dysfunction. Am J PhysiolHeart Circ Physiol 2005;289:H1744–H1751

15. Kumar SD, Yong SK, Dheen ST, Bay BH, Tay SSW. Cardiac malformationsare associated with altered expression of vascular endothelial growthfactor and endothelial nitric oxide synthase genes in embryos of diabeticmice. Exp Biol Med (Maywood) 2008;233:1421–1432

16. Martinsen BJ. Reference guide to the stages of chick heart embryology.Dev Dyn 2005;233:1217–1237

17. Hove JR, Köster RW, Forouhar AS, Acevedo-Bolton G, Fraser SE, GharibM. Intracardiac fluid forces are an essential epigenetic factor for embry-onic cardiogenesis. Nature 2003;421:172–177

18. Clark EB, Rosenquist GC. Spectrum of cardiovascular anomalies followingcardiac loop constriction in the chick embryo. Birth Defects Orig Artic Ser1978;14:431–442

19. Liu A, Wang R, Thornburg KL, Rugonyi S. Efficient postacquisition syn-chronization of 4-D nongated cardiac images obtained from optical co-herence tomography: application to 4-D reconstruction of the chickembryonic heart. J Biomed Opt 2009;14:044020

20. Rothenberg F, Fisher SA, Watanabe M. Sculpting the cardiac outflow tract.Birth Defects Res C Embryo Today 2003;69:38–45

21. Hamburger V, Hamilton HL. A series of normal stages in the developmentof the chick embryo. 1951. Dev Dyn 1992;195:231–272

22. Pentecost JO, Silva C, Pesticelli MJ Jr, Thornburg KL. Modeling cardio-genesis: the challenges and promises of 3D reconstruction. Curr Top DevBiol 2003;56:115–143

23. Miller RR, Burum AL, Leithard ME, Hart JD. Hyperglycemia-inducedchanges in hepatic membrane fatty acid composition correlate with in-creased caspase-3 activities and reduced chick embryo viability. CompBiochem Physiol B Biochem Mol Biol 2005;141:323–330

24. Ahnfelt-Rønne J, Jørgensen MC, Hald J, Madsen OD, Serup P, Hecksher-Sørensen J. An improved method for three-dimensional reconstruction ofprotein expression patterns in intact mouse and chicken embryos andorgans. J Histochem Cytochem 2007;55:925–930

25. Gebhardt DOE. The teratogenic action of propylene glycol (propanediol-1,2) and propanediol-1,3 in the chick embryo. Teratology 1968;1:153–161

26. Kalter H. Teratology in the 20th century: environmental causes of con-genital malformations in humans and how they were established. Neuro-toxicol Teratol 2003;25:131–282

27. Mizisin AP, Li L, Perello M, et al. Polyol pathway and osmoregulation inJS1 Schwann cells grown in hyperglycemic and hyperosmotic conditions.Am J Physiol 1996;270:F90–F97

28. Carver FM, Shibley IA Jr, Pennington JS, Pennington SN. Differential ex-pression of glucose transporters during chick embryogenesis. Cell Mol LifeSci 2001;58:645–652

29. Moley KH, Chi MM, Mueckler MM. Maternal hyperglycemia alters glucosetransport and utilization in mouse preimplantation embryos. Am J Physiol1998;275:E38–E47

30. Moley KH. Diabetes and preimplantation events of embryogenesis. SeminReprod Endocrinol 1999;17:137–151

31. Heilig CW, Saunders T, Brosius FC 3rd, et al. Glucose transporter-1-deficientmice exhibit impaired development and deformities that are similar to di-abetic embryopathy. Proc Natl Acad Sci U S A 2003;100:15613–15618

32. Svensson AM, Borg LA, Eriksson UJ. Glucose metabolism in embryos ofnormal and diabetic rats during organogenesis. Acta Endocrinol (Copenh)1992;127:252–257

33. Pedersen JF, Mølsted-Pedersen L. Early growth retardation in diabeticpregnancy. BMJ 1979;1:18–19

34. Datar S, Bhonde RR. Shell-less chick embryo culture as an alternative invitro model to investigate glucose-induced malformations in mammalianembryos. Rev Diabet Stud 2005;2:221–227

35. Reis LH, Gama LT, Soares MC. Effects of short storage conditions andbroiler breeder age on hatchability, hatching time, and chick weights.Poult Sci 1997;76:1459–1466

36. Becerra JE, Khoury MJ, Cordero JF, Erickson JD. Diabetes mellitus duringpregnancy and the risks for specific birth defects: a population-based case-control study. Pediatrics 1990;85:1–9

37. Schaefer-Graf UM, Buchanan TA, Xiang A, Songster G, Montoro M, KjosSL. Patterns of congenital anomalies and relationship to initial maternalfasting glucose levels in pregnancies complicated by type 2 and gestationaldiabetes. Am J Obstet Gynecol 2000;182:313–320

38. Suzuki N, Svensson K, Eriksson UJ. High glucose concentration inhibitsmigration of rat cranial neural crest cells in vitro. Diabetologia 1996;39:401–411

39. Molin DGM, Roest PAM, Nordstrand H, et al. Disturbed morphogenesis ofcardiac outflow tract and increased rate of aortic arch anomalies in theoffspring of diabetic rats. Birth Defects Res A Clin Mol Teratol 2004;70:927–938

40. Zhao Z. Cardiac malformations and alteration of TGFbeta signaling systemin diabetic embryopathy. Birth Defects Res B Dev Reprod Toxicol 2010;89:97–105

41. Mizutani M, Okuda Y, Yamaoka T, et al. High glucose and hyperosmolarityincrease platelet-derived growth factor mRNA levels in cultured humanvascular endothelial cells. Biochem Biophys Res Commun 1992;187:664–669

42. Rosso A, Balsamo A, Gambino R, et al. p53 Mediates the accelerated onsetof senescence of endothelial progenitor cells in diabetes. J Biol Chem2006;281:4339–4347

43. Chang BD, Watanabe K, Broude EV, et al. Effects of p21Waf1/Cip1/Sdi1 oncellular gene expression: implications for carcinogenesis, senescence, andage-related diseases. Proc Natl Acad Sci U S A 2000;97:4291–4296

44. Chattergoon NN, Giraud GD, Thornburg KL. Thyroid hormone inhibitsproliferation of fetal cardiac myocytes in vitro. J Endocrinol 2007;192:R1–R8

45. Chattergoon NN, Giraud GD, Louey S, Stork P, Fowden AL, Thornburg KL.Thyroid hormone drives fetal cardiomyocyte maturation. FASEB J 2012;26:397–408

46. de Nadal E, Alepuz PM, Posas F. Dealing with osmostress through MAPkinase activation. EMBO Rep 2002;3:735–740

47. Moley KH. Hyperglycemia and apoptosis: mechanisms for congenitalmalformations and pregnancy loss in diabetic women. Trends EndocrinolMetab 2001;12:78–82

48. Dhanasekaran N, Wu YK, Reece EA. Signaling pathways and diabeticembryopathy. Semin Reprod Endocrinol 1999;17:167–174

HYPERGLYCEMIA SUPPRESSES EMBRYONIC DEVELOPMENT

242 DIABETES, VOL. 62, JANUARY 2013 diabetes.diabetesjournals.org